Dependency Parsing On Universal Dependencies
Metrics
LAS
UAS
Results
Performance results of various models on this benchmark
Model Name | LAS | UAS | Paper Title | Repository |
|---|---|---|---|---|
| UDify | 80.43 | 85.69 | 75 Languages, 1 Model: Parsing Universal Dependencies Universally | - |
| UDPipe 2.0 + mBERT + FLAIR | 84.60 | 87.64 | Evaluating Contextualized Embeddings on 54 Languages in POS Tagging, Lemmatization and Dependency Parsing | - |
| TurkuNLP | 73.28 | 60.99 | Turku Neural Parser Pipeline: An End-to-End System for the CoNLL 2018 Shared Task | - |
| UDPipe 2.0 | 73.11 | 61.25 | UDPipe 2.0 Prototype at CoNLL 2018 UD Shared Task | - |
| HIT-SCIR | 75.84 | - | Towards Better UD Parsing: Deep Contextualized Word Embeddings, Ensemble, and Treebank Concatenation | - |
| Stanford+ | 74.16 | 62.08 | Universal Dependency Parsing from Scratch | - |
0 of 6 row(s) selected.